280 Chamberlin and Leverett — Studies of the 



This last statement requires qualification, for, if the joining of 

 the basins by the reversal of streams is true, then we must 

 either suppose that the cols between these basins had been de- 

 stroyed by the backward working of the stream (which would 

 be likely to be true only in a case of approximate base-level- 

 ling), or we must suppose that the cols dividing these basins 

 were trenched to a depth in rock equal to the rock cutting 

 under the first hypothesis. The deep trenches of the Alle- 

 gheny with their steep sides, show that the region was in the 

 early stages of developing a new valley system, and that, 

 therefore, the cols were presumably intact up to the heights 

 of the old gradation-plane, at least, which, under this hypoth- 

 esis, is referred to the Tertiary period. While this hypothesis 

 escapes a large part of the rock cutting by referring it to pre- 

 glacial times, it greatly magnifies the first glacial epoch. 

 When we consider the limited area of the glaciated region be- 

 tween the edge of the drift and the descent to the Lake Erie 

 valley (which limited the amount of inwash after the ice began 

 to retire), we realize the import of this hypothesis. It means 

 that some 300 feet of drift filled the valley for 250 miles at 

 least, while the ice edge occupied the narrow belt between the 

 glacial boundary and the basin of Lake Erie. 



Hypothesis III. — This hypothesis postulates a limited 

 amount of trenching of the old Tertiary gradation-plane before 

 the ice reached the basin, the amount of this trenching being- 

 inferred from the size of buried trenches found in the bottoms 

 of the upper Allegheny and Conewango basins, which, the 

 theory assumes, were buried by the incoming ice without much 

 modification and have failed to suffer any of moment since. 

 The evidence of such trenches is rather meagre, but a small 

 proportion of the wells in the broad valley of the Allegheny- 

 above the Kinzua col and in the Conewango and Brokenstraw 

 valleys sink below the majority of wells to depths amounting 

 to 75 to 100 feet. Wells not far distant from each other some- 

 times show this difference, from which it is inferred that the 

 sides of the trench are abrupt. The fact that only a minor 

 percentage, perhaps one in ten, of the wells in these basins 

 enter this deeper portion, seems to indicate that it was only 

 the beginning of a process of deeper erosion and had not pro- 

 gressed far. These trenches do not seem to be at all compar- 

 able to the trenches on the lower Allegheny below the grada- 

 tion-plane. The hypothesis recognizes the fact that these 

 northern preglacial streams could reach lower levels by shorter 

 routes than the more southerly ones, even if the latter at length 

 emptied into the Lake Erie basin, and that, therefore, the 

 trenches found here are the greatest that legitimate interpre- 

 tation could appeal to. Trenches in the lower section would 

 probably be smaller. 



